Re: [RFC] Revert "drm/sched: Split free_job into own work item"

[Mario Limonciello <mario.limonciello@xxxxxxx>]

On 1/22/2024 20:11, Mario Limonciello wrote:
> commit f7fe64ad0f22 ("drm/sched: Split free_job into own work item")
> causes graphics hangs at GDM or right after logging in on a
> Framework 13 AMD laptop (containing a Phoenix APU).
>
> This reverts commit f7fe64ad0f22ff034f8ebcfbd7299ee9cc9b57d7.
>
> Fixes: f7fe64ad0f22 ("drm/sched: Split free_job into own work item")
> Signed-off-by: Mario Limonciello <mario.limonciello@xxxxxxx>
> ---
> This is a regression introduced in 6.8-rc1, bisected from 6.7.
> This revert done on top of 6.8-rc1 fixes the issue.
>
> I'm happy to gather any data to use to properly debug if that is
> preferable to a revert.

BTW - nothing in kernel log when this happens.
I did catch an ftrace with all the GPU scheduler trace events though:

https://gist.github.com/superm1/1d3fc752a414f324fbde526052573acd

> ---
>   drivers/gpu/drm/scheduler/sched_main.c | 133 +++++++++----------------
>   include/drm/gpu_scheduler.h            |   4 +-
>   2 files changed, 48 insertions(+), 89 deletions(-)
>
> diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c
> index 550492a7a031..91c96ab53fb5 100644
> --- a/drivers/gpu/drm/scheduler/sched_main.c
> +++ b/drivers/gpu/drm/scheduler/sched_main.c
> @@ -369,32 +369,6 @@ static void drm_sched_run_job_queue(struct drm_gpu_scheduler *sched)
>   		queue_work(sched->submit_wq, &sched->work_run_job);
>   }
>   
> -/**
> - * __drm_sched_run_free_queue - enqueue free-job work
> - * @sched: scheduler instance
> - */
> -static void __drm_sched_run_free_queue(struct drm_gpu_scheduler *sched)
> -{
> -	if (!READ_ONCE(sched->pause_submit))
> -		queue_work(sched->submit_wq, &sched->work_free_job);
> -}
> -
> -/**
> - * drm_sched_run_free_queue - enqueue free-job work if ready
> - * @sched: scheduler instance
> - */
> -static void drm_sched_run_free_queue(struct drm_gpu_scheduler *sched)
> -{
> -	struct drm_sched_job *job;
> -
> -	spin_lock(&sched->job_list_lock);
> -	job = list_first_entry_or_null(&sched->pending_list,
> -				       struct drm_sched_job, list);
> -	if (job && dma_fence_is_signaled(&job->s_fence->finished))
> -		__drm_sched_run_free_queue(sched);
> -	spin_unlock(&sched->job_list_lock);
> -}
> -
>   /**
>    * drm_sched_job_done - complete a job
>    * @s_job: pointer to the job which is done
> @@ -414,7 +388,7 @@ static void drm_sched_job_done(struct drm_sched_job *s_job, int result)
>   	dma_fence_get(&s_fence->finished);
>   	drm_sched_fence_finished(s_fence, result);
>   	dma_fence_put(&s_fence->finished);
> -	__drm_sched_run_free_queue(sched);
> +	drm_sched_run_job_queue(sched);
>   }
>   
>   /**
> @@ -1092,10 +1066,8 @@ drm_sched_get_finished_job(struct drm_gpu_scheduler *sched)
>   						typeof(*next), list);
>   
>   		if (next) {
> -			if (test_bit(DMA_FENCE_FLAG_TIMESTAMP_BIT,
> -				     &next->s_fence->scheduled.flags))
> -				next->s_fence->scheduled.timestamp =
> -					dma_fence_timestamp(&job->s_fence->finished);
> +			next->s_fence->scheduled.timestamp =
> +				dma_fence_timestamp(&job->s_fence->finished);
>   			/* start TO timer for next job */
>   			drm_sched_start_timeout(sched);
>   		}
> @@ -1145,29 +1117,7 @@ drm_sched_pick_best(struct drm_gpu_scheduler **sched_list,
>   EXPORT_SYMBOL(drm_sched_pick_best);
>   
>   /**
> - * drm_sched_free_job_work - worker to call free_job
> - *
> - * @w: free job work
> - */
> -static void drm_sched_free_job_work(struct work_struct *w)
> -{
> -	struct drm_gpu_scheduler *sched =
> -		container_of(w, struct drm_gpu_scheduler, work_free_job);
> -	struct drm_sched_job *job;
> -
> -	if (READ_ONCE(sched->pause_submit))
> -		return;
> -
> -	job = drm_sched_get_finished_job(sched);
> -	if (job)
> -		sched->ops->free_job(job);
> -
> -	drm_sched_run_free_queue(sched);
> -	drm_sched_run_job_queue(sched);
> -}
> -
> -/**
> - * drm_sched_run_job_work - worker to call run_job
> + * drm_sched_run_job_work - main scheduler thread
>    *
>    * @w: run job work
>    */
> @@ -1176,50 +1126,64 @@ static void drm_sched_run_job_work(struct work_struct *w)
>   	struct drm_gpu_scheduler *sched =
>   		container_of(w, struct drm_gpu_scheduler, work_run_job);
>   	struct drm_sched_entity *entity;
> -	struct dma_fence *fence;
> -	struct drm_sched_fence *s_fence;
> -	struct drm_sched_job *sched_job;
> +	struct drm_sched_job *cleanup_job;
>   	int r;
>   
>   	if (READ_ONCE(sched->pause_submit))
>   		return;
>   
> +	cleanup_job = drm_sched_get_finished_job(sched);
>   	entity = drm_sched_select_entity(sched);
> -	if (!entity)
> -		return;
>   
> -	sched_job = drm_sched_entity_pop_job(entity);
> -	if (!sched_job) {
> -		complete_all(&entity->entity_idle);
> +	if (!entity && !cleanup_job)
>   		return;	/* No more work */
> -	}
>   
> -	s_fence = sched_job->s_fence;
> +	if (cleanup_job)
> +		sched->ops->free_job(cleanup_job);
>   
> -	atomic_add(sched_job->credits, &sched->credit_count);
> -	drm_sched_job_begin(sched_job);
> +	if (entity) {
> +		struct dma_fence *fence;
> +		struct drm_sched_fence *s_fence;
> +		struct drm_sched_job *sched_job;
> +
> +		sched_job = drm_sched_entity_pop_job(entity);
> +		if (!sched_job) {
> +			complete_all(&entity->entity_idle);
> +			if (!cleanup_job)
> +				return;	/* No more work */
> +			goto again;
> +		}
>   
> -	trace_drm_run_job(sched_job, entity);
> -	fence = sched->ops->run_job(sched_job);
> -	complete_all(&entity->entity_idle);
> -	drm_sched_fence_scheduled(s_fence, fence);
> +		s_fence = sched_job->s_fence;
>   
> -	if (!IS_ERR_OR_NULL(fence)) {
> -		/* Drop for original kref_init of the fence */
> -		dma_fence_put(fence);
> +		atomic_inc(&sched->credit_count);
> +		drm_sched_job_begin(sched_job);
>   
> -		r = dma_fence_add_callback(fence, &sched_job->cb,
> -					   drm_sched_job_done_cb);
> -		if (r == -ENOENT)
> -			drm_sched_job_done(sched_job, fence->error);
> -		else if (r)
> -			DRM_DEV_ERROR(sched->dev, "fence add callback failed (%d)\n", r);
> -	} else {
> -		drm_sched_job_done(sched_job, IS_ERR(fence) ?
> -				   PTR_ERR(fence) : 0);
> +		trace_drm_run_job(sched_job, entity);
> +		fence = sched->ops->run_job(sched_job);
> +		complete_all(&entity->entity_idle);
> +		drm_sched_fence_scheduled(s_fence, fence);
> +
> +		if (!IS_ERR_OR_NULL(fence)) {
> +			/* Drop for original kref_init of the fence */
> +			dma_fence_put(fence);
> +
> +			r = dma_fence_add_callback(fence, &sched_job->cb,
> +						   drm_sched_job_done_cb);
> +			if (r == -ENOENT)
> +				drm_sched_job_done(sched_job, fence->error);
> +			else if (r)
> +				DRM_DEV_ERROR(sched->dev, "fence add callback failed (%d)\n",
> +					  r);
> +		} else {
> +			drm_sched_job_done(sched_job, IS_ERR(fence) ?
> +					   PTR_ERR(fence) : 0);
> +		}
> +
> +		wake_up(&sched->job_scheduled);
>   	}
>   
> -	wake_up(&sched->job_scheduled);
> +again:
>   	drm_sched_run_job_queue(sched);
>   }
>   
> @@ -1304,7 +1268,6 @@ int drm_sched_init(struct drm_gpu_scheduler *sched,
>   	atomic_set(&sched->credit_count, 0);
>   	INIT_DELAYED_WORK(&sched->work_tdr, drm_sched_job_timedout);
>   	INIT_WORK(&sched->work_run_job, drm_sched_run_job_work);
> -	INIT_WORK(&sched->work_free_job, drm_sched_free_job_work);
>   	atomic_set(&sched->_score, 0);
>   	atomic64_set(&sched->job_id_count, 0);
>   	sched->pause_submit = false;
> @@ -1432,7 +1395,6 @@ void drm_sched_wqueue_stop(struct drm_gpu_scheduler *sched)
>   {
>   	WRITE_ONCE(sched->pause_submit, true);
>   	cancel_work_sync(&sched->work_run_job);
> -	cancel_work_sync(&sched->work_free_job);
>   }
>   EXPORT_SYMBOL(drm_sched_wqueue_stop);
>   
> @@ -1445,6 +1407,5 @@ void drm_sched_wqueue_start(struct drm_gpu_scheduler *sched)
>   {
>   	WRITE_ONCE(sched->pause_submit, false);
>   	queue_work(sched->submit_wq, &sched->work_run_job);
> -	queue_work(sched->submit_wq, &sched->work_free_job);
>   }
>   EXPORT_SYMBOL(drm_sched_wqueue_start);
> diff --git a/include/drm/gpu_scheduler.h b/include/drm/gpu_scheduler.h
> index 5acc64954a88..72f64bd0a6eb 100644
> --- a/include/drm/gpu_scheduler.h
> +++ b/include/drm/gpu_scheduler.h
> @@ -495,10 +495,9 @@ struct drm_sched_backend_ops {
>    *                 waits on this wait queue until all the scheduled jobs are
>    *                 finished.
>    * @job_id_count: used to assign unique id to the each job.
> - * @submit_wq: workqueue used to queue @work_run_job and @work_free_job
> + * @submit_wq: workqueue used to queue @work_run_job
>    * @timeout_wq: workqueue used to queue @work_tdr
>    * @work_run_job: work which calls run_job op of each scheduler.
> - * @work_free_job: work which calls free_job op of each scheduler.
>    * @work_tdr: schedules a delayed call to @drm_sched_job_timedout after the
>    *            timeout interval is over.
>    * @pending_list: the list of jobs which are currently in the job queue.
> @@ -528,7 +527,6 @@ struct drm_gpu_scheduler {
>   	struct workqueue_struct		*submit_wq;
>   	struct workqueue_struct		*timeout_wq;
>   	struct work_struct		work_run_job;
> -	struct work_struct		work_free_job;
>   	struct delayed_work		work_tdr;
>   	struct list_head		pending_list;
>   	spinlock_t			job_list_lock;